• 中国核心期刊数据库收录期刊
  • 中文科技期刊数据库收录期刊
  • 中国期刊全文数据库收录期刊
  • 中国学术期刊综合评价数据库统计源期刊等

快速检索引用检索图表检索高级检索

中国医药导刊 ›› 2026, Vol. 28 ›› Issue (5): 558-558-567.doi: 10.1009-0959.2026.050026

• 基础研究 • 上一篇    下一篇

基于生物信息学与机器学习分析重度抑郁症相关铁死亡关键基因

刘继俊, 张慧*   

  1. 安徽中医药大学针灸推拿学院,安徽 合肥 230012
  • 收稿日期:2025-09-10 修回日期:2026-03-26 接受日期:2026-04-20 出版日期:2026-05-28 发布日期:2026-06-29
  • 基金资助:
    安徽省自然科学基金项目(2208085MH272);安徽省高校自然科学研究重点项目(2023AH050800)

Bioinformatics and Machine Learning-Based Analysis of Key Ferroptosis-Related Genes in Major Depressive Disorder

LIU Jijun, ZHANG Hui*   

  1. College of Acupuncture-Moxibustion and Tuina Anhui University of Chinese Medicine Anhui Hefei 230012, China
  • Received:2025-09-10 Revised:2026-03-26 Accepted:2026-04-20 Online:2026-05-28 Published:2026-06-29

摘要:

目的:基于生物信息学与机器学习技术筛选重度抑郁症(major depressive disorderMDD)相关铁死亡关键基因并构建诊断模型,为MDD的临床辅助诊断提供新型生物标志物。方法:从GEO基因数据库中获得MDD数据集,经R软件标准化和批次效应校正后,筛选差异基因(DEGs),将DEGsFerrDb数据库铁死亡基因集取交集,得到MDD铁死亡相关基因(FRGs),对FRGs进行GOKEGG富集分析。随后采用支持向量机(SVM)、随机森林(RF)和最小绝对收敛和选择算子(LASSO)回归多算法联合筛选MDD铁死亡相关关键基因(Hub-FRGs),并构建逻辑回归诊断模型。通过受试者工作特征曲线(ROC)评估模型效能,在独立验证集中进行验证,再进行免疫浸润分析。结果:获得29FRGsGO/KEGG主要富集于氧化应激、炎症、代谢等相关生物功能与通路。经机器学习算法筛选出MAFGLCN2G6PDHub-FRGs。所构建诊断模型在训练集曲线下面积(AUC)达0.932,在GSE76826AUC=0.867)验证集中表现良好。免疫浸润结果显示巨噬细胞、CD4⁺ TCD8⁺ T细胞与Hub-FRGs存在相关性。结论:铁死亡关键基因 MAFGLCN2G6PD具备作为MDD诊断标志物的潜力,可为MDD的临床诊断与治疗提供新的研究思路。


关键词: 生物信息学, 机器学习, 抑郁症, 铁死亡

Abstract:

Objective: To identify key ferroptosis-related genes FRGs in major depressive disorder MDD and construct a diagnostic model using bioinformatics and machine learning aiming to provide novel biomarkers for the clinical auxiliary diagnosis of MDD.Methods: MDD datasets were obtained from the GEO database. After standardization and batch effect correction differentially expressed genes DEGs were screened. The intersection of DEGs with the ferroptosis-related gene set from FerrDb defined MDD-specific FRGs which underwent functional enrichment analysis GO and KEGG. Hub-FRGs were subsequently identified through the combined application of support vector machine SVM), random forest RF), and least absolute shrinkage and selection operator LASSO regression. A logistic regression diagnostic model was constructed and evaluated using receiver operating characteristic ROC curve analysis followed by validation in an independent set and immune infiltration analysis.Results: Twenty-nine FRGs were obtained primarily enriched in biological functions and pathways related to oxidative stress inflammation and metabolism. Three genesMAFG LCN2 and G6PDwere selected as Hub-FRGs. The diagnostic model achieved an AUC of 0.932 in the training set and an AUC of 0.867 in the GSE76826 validation set. Immune infiltration analysis revealed significant correlations of macrophages CD4⁺ T cells and CD8⁺ T cells with the Hub-FRGs.Conclusion: The identified ferroptosis-related hub genes MAFG LCN2 and G6PD possess the potential to serve as diagnostic biomarkers for MDD providing new insights for the diagnosis and treatment of MDD.


Key words: Bioinformatics , Machine learning , Depression , Ferroptosis

中图分类号: